Reducing Rapid Movements

[combatraffi 0]

Hey Guys,

I am tyring to print off a ball catch with an UM2. As high res as possible as there will be 2 concentric members that will need to slide against eachother. Generally speaking the resolution is excellent when you go to .05 layers slow down etc. But I keep getting rapid movement where the machine will complete one revolution of the thin walled cylinder and then fly to another section and start the next layer. This leaves a little glob on that spot completely killing the finish. this happens on most layers. It looks ALOT like the pin cylinder(?) on a music box. I tried spiralize but that produced some seriously broken slices... Any ideas?

 

[gr5 2,172]

Picture please.

Usually you *want* it to go faster when it isn't extruding. That's the "travel speed" which you can change but usually works best at 150mm/sec minimum. 200 to 300 works even better and tends to leave less blobs.

Instead you probably want to enable retraction. There are 6 retraction settings in the 3 different settings menus. Make sure it is checked and in the "expert settings" section change both distances to 0 mm.

Without retraction you get what we call "stringing". Strings that connect parts that weren't meant to be connected. You also often get blobs at the start and/or end of the strings.

Or maybe you aren't talking about stringing at all! Please post a picture.

 

[peggyb 170]

i think you mean the z-scar, which seams to be back in Cura 13.11.

For some reason the nozzle moves outside the layer, leaving a scar that is visible along the whole surface.

It was not present in 13.10

 

[Dim3nsioneer 557]

i think you mean the z-scar, which seams to be back in Cura 13.11.

For some reason the nozzle moves outside the layer, leaving a scar that is visible along the whole surface.

It was not present in 13.10

 

Thank you for the explanation... I really was thinking very hard why I suddenly get these blobs...

@Daid: Is it possible to fix that for the next version, please (if not done already... :wink: )?

 

[gr5 2,172]

Before you claim it's the z scar please back that up with evidence. Look at the finished printed part and at the same time look at it in Cura in layer view and locate where the z axis moves up a layer. Actually it's MUCH easier to see in repetier host. If you install repetier host (it's free) and then simply run it and drag and drop the gcode onto RH you can them pan, zoom, rotate around the part and see where the z movements occur.

 

[Dim3nsioneer 557]

Before you claim it's the z scar please back that up with evidence. [...]

 

You're absolutely right. I'll have a closer look into that issue today or tomorrow.

 

[Dim3nsioneer 557]

Here comes the evidence... but not for the exact effect described by PeggyB...

I analyzed this print:

 

As you can see in the center of the image, there is a knob produced for five layers towards the top of the print. I checked the layers in Cura 13.11.2 and with the GCODE-viewer available at gcode.ws (as it is not well visible in Cura as pointed out by gr5):

 

I found the artefact to be visible also in Cura 13.11 and 13.10 when using the same settings. However, I also found out that the knob disappears if I change the shell thickness from 1.2mm down to 0.8mm:

 

 

From that point on, I think, a Cura insider is needed for further investigation...

 

[peggyb 170]

here is my backup:

straight print, no fill, just walls

13.09 and 13.10 are giving the same results.

13.11 also effects the inside, as this being a usb-stick cover, it didn't fit in there. The 13.09 did fit perfectly.

 

[illuminarti 18]

I think Dim3nsioneer's print issue is caused by infill protruding through the outer wall. It looks like the wall is thin, so there's not really room to do a 1.2mm thick shell on both sides. Stepping down to 0.8mm leaves more room inside to neatly fill between the walls.

Depending on the angle that the fill meets that wall, you can get some odd effects. Reducing the infill overlap percentage a bit might help also.

 

[Dim3nsioneer 557]

I think Dim3nsioneer's print issue is caused by infill protruding through the outer wall. It looks like the wall is thin, so there's not really room to do a 1.2mm thick shell on both sides. Stepping down to 0.8mm leaves more room inside to neatly fill between the walls.

Depending on the angle that the fill meets that wall, you can get some odd effects. Reducing the infill overlap percentage a bit might help also.

 

That would have been too easy... :wink:

The original wall thickness was 3mm. Thus, the infill gets 3mm-2x1.2mm=0.6mm space; with the decrased shell thickness, it gets 0.8mm more.

So I did a test, inreasing the wall thickness to 3.8mm, giving the additional space to the infill. This is the result:

 

The knob is still there, at the very same spot as before...

 

I cannot decrease the infill overlap as I have it already set to zero (and no, I do not have gaps or not-sticking issues as my Ultimaker is quite well adjusted). Wouldn't it anyway be a bit strange if the slicer would produce an effect from the infill overlap? I would rather expect the printer produce a knob due to that as there would be just to much material. But I'm no slicer-expert...

 

[illuminarti 18]

Weird. Can you post the gcode somewhere? I'd like to take a look at it.

 

[Dim3nsioneer 557]

Weird. Can you post the gcode somewhere? I'd like to take a look at it.

 

You'll find it here. The knob is visible from layer 228 to layer 232. Please don't worry about the settings; it's Cura's default (apart from the shell thickness; the gcode was generated with these settings).

 

[illuminarti 18]

Are you absolutely sure there's not something odd about your mesh at that point? Looking at the gcode, as it traces around the outer perimeter, it jumps out about 0.2mm, and then back to exactly where it started, before continuing around. Indeed, it's not related to infill or anything about the mechanics of the print; it's very deliberately deciding to print a little bump at that point, as if there was something there in the model to print. The only slightly odd thing about it is that the bump retraces its steps, starting and ending in exactly the same spot.

My guess would be that there's a stray or flipped triangle there - something confusing the slicer.

 

[gr5 2,172]

This is a fascinating topic! Please post the STL file.

 

[Dim3nsioneer 557]

hmmm... The model was made with Blender... which caused me other troubles before with Curaengine.

However, I checked the normals, they seem ok from what I saw. Sometimes Blender shifts a point a bit (just by a few microns!).

The circle was made with 60 segments. If one of the 60 outermost points is shifted just a bit, then it might confuse Curaengine. But then I would expect that the artefact appears independently of the shell thickness setting.

Feel free to check the https://dl.dropboxusercontent.com/u/15273556/ultimaker/gcode/knob%20-%20example.stl.

 

[gr5 2,172]

I think it's mostly a bug in Cura. I've seen when you set for example to 50% overlap sometimes it does 50% in X or Y only but ends up being 100% in the other axis and similarly strange things. The way Cuya calculates this overlap with the shell seems to have a bug. I think the bug shows up in that exact spot because your STL is .2mm wider there.

 

[billdempsey 0]

I have had one or two models that I've printed which exhibited this sort of strange bump, too. If I slice them with another slicer, the bump usually disappears. The vast majority of models slice really well in Cura, though. I much prefer it over anything else I've tried.